AI agent adoption is accelerating quickly. Gartner predicts that by 2028, the average global Fortune 500 enterprise could have more than 150,000 agents in use, up from fewer than 15 in 2025. Gartner also reports that only 13% of organizations say they have the right AI agent governance in place.1

It’s getting harder to know which agents already exist within an organization, who owns them, and who can access them. Business users may struggle to find the right agent for a task, while IT and AI teams need visibility into how those agents connect to enterprise data, systems, and tools.

Oracle AI Data Platform addresses this with AI Registry, which gives organizations a central way to register, manage, and apply access controls to approved AI agents and supporting capabilities.

More Agents, More Complexity

As organizations create and use more AI agents, they need to make a growing number of decisions about what to build, what to reuse, what to govern, where AI can help reduce repetitive work, and how approved capabilities should be made available across the business.

Those decisions depend on answering questions such as:

  • Which agents and AI capabilities already exist?
  • Have tools been reviewed and approved for broader use?
  • What common functionality can be reused across teams?
  • Where are teams still doing repetitive manual work that could be supported by an agent, tool, or reusable AI capability?
  • Who owns and maintains each capability?
  • Who can access each capability?

Without clear answers, organizations risk creating unnecessary agent sprawl, rebuilding capabilities that already exist, and making it harder for business users to find the right support. IT and AI teams also need visibility into which agents are being used, what data and systems they can access, and whether that access aligns with enterprise permissions and governance policies.

AI Registry: Visibility and Access Management for IT and AI Teams

Oracle AI Data Platform AI Registry gives IT, AI platform teams, and administrators a central place to manage registered AI capabilities and configure access controls.

Administrators can register agents and MCP servers, including agents from Oracle AI Data Platform, Fusion, Database, and other supported environments, as well as compatible external agents that support the Agent2Agent (A2A) protocol.

A2A agent cards provide details such as supported skills, streaming capabilities, and input and output types, while MCP server metadata includes names and descriptions. This information helps administrators understand the capabilities available through AI Registry.

This visibility also helps teams discover and reuse existing capabilities rather than build duplicate agents or tools.

AI Registry provides a centralized view of registered AI agents and MCP servers, including their metadata and configured access controls.

How AI Registry Works in Oracle AI Data Platform

As part of Oracle AI Data Platform, AI Registry fits into a broader workflow for developing and using enterprise AI. Organizations can prepare and connect data, build or bring in agents, and make approved capabilities available to users through supported applications and experiences.

Organizations remain responsible for establishing and implementing their own AI governance policies and practices, including determining which capabilities are approved for use and who should be permitted to access them.

Prepare and connect data

Data teams make relevant enterprise data available and add the business context AI needs to help interpret it. Oracle AI Data Platform helps represent business meaning, relationships, and semantics so agents can reason over enterprise data in context.

Build AI agents

Developers and other builders create agents in Oracle AI Data Platform and other supported environments. They can apply reusable skills to give specialized agents task-specific guidance, business rules, and procedures. Compatible agents from supported Oracle and third-party environments can also be registered in AI Registry, including agents that support the A2A protocol where applicable.

Register and manage agents

Administrators use AI Registry to register approved agents and MCP servers from across the organization, define and configure Role Based Access Control on the agents and MCP servers.

Give users access to relevant agents

Organizations can use role-based access control in AI Registry to manage access to registered agents and MCP servers. Based on their roles and permissions, users can access approved agents and MCP tools through Oracle Agent Hub, compatible, supported third-party AI assistants, and other applications and business workflows.

Oracle Agent Hub is Oracle AI Data Platform’s business-user experience for working with approved enterprise agents. It can help users identify and coordinate the right agent or agents for a task, while AI Registry provides a centralized way to manage those capabilities.

Benefits for IT, AI Teams, and Business Users

For IT and AI teams, AI Registry provides a central view of approved agents and supporting capabilities. Teams can manage access and organize what’s available rather than managing each registered capability separately.

For developers and other builders, the registry makes it easier to see what already exists. That creates more opportunities to reuse approved agents, tools, and connections rather than rebuilding the same capabilities for each project.

Business users, can more easily find approved AI capabilities through the experiences they already use, whether that’s Oracle Agent Hub, a supported AI assistant, or a business application.
This helps organizations expand access to enterprise AI while maintaining visibility over the agents and AI capabilities their business uses.

Get Started with AI Registry

As organizations expand their use of AI agents, they need a way to understand what exists, control who can use it, and make approved capabilities available across the business. AI Registry provides a centralized way to register approved AI capabilities and configure access controls while giving organizations flexibility in how employees access and use enterprise AI.

Together with an organization’s own governance policies and practices, these capabilities can support the management and reuse of enterprise AI agents and tools.

Learn More

1Source: Gartner, “Gartner Identifies Six Steps to Manage AI Agent Sprawl,” April 28, 2026.